Though markets generally are reasonably efficient, there are instances when changes in the price of an asset don’t necessarily reflect the available information. Such instances are called market anomalies. However, such inefficiencies should be present over a relatively period of time to be classified as a market anomaly.
Data mining (or data snooping) is the process of analyzing data with the motive of developing a hypothesis. It is different from typical research work where hypothesis based on economic rationale is first prepared and then data is analyzed to approve/disprove the hypothesis. To avoid data mining bias, researchers must make sure that economic rationale exists between the variables and that the relationship holds true for different large enough data sample.
We now discuss some of the well-known market anomalies:
- Time-series anomalies which are further classified into:
- Calendar anomalies: Some anomalies are specific to certain time periods. For e.g. January effect or turn-of-the-month anomaly refers to the trend of small market capitalization stocks performing relatively well during first five trading days of January vs. rest of the year. Various explanations have been offered to explain this effect. One is related to investors selling highly volatile small cap stocks in losses during December to reduce tax liabilities. The investor then buys back these securities in January to increase potential returns. Another explanation refers to window dressing in which portfolio managers sell riskier assets at the end of the year to reduce the risk profile of their funds in the annual reports. The securities are then bought back in January to generate potential above-market Similarly, day-of-the-week anomaly refers to market returns being generally lowest on Monday than the other four days of the week. Weekend effect refers to returns being lower on weekends than on weekdays. Similarly, holiday effect refers to returns being higher on days prior to holidays than other days.
- Momentum and overreaction anomalies are related to stock reaction to the release of unexpected public information. Researchers have argued that stocks tend to “overreact” to the release of such information. Thus stocks may become overpriced (or underpriced) on positive (or negative) news flow. This is known as overreaction effect. A trading strategy based on this anomaly entailed buying (or selling) stocks with consistently low returns (or higher returns) over a three to five year period as they tend to outperform the broader market in the subsequent period. Similarly, momentum effect refers to a stock (or any other asset class) that has risen (or fallen) a lot in the short term tends to carry on with this gain (or loss) momentum in the near to medium term. This effect challenges the existence of a weak form of market efficiency as investors tend to make abnormal returns based on past information.
- Cross-sectional anomalies:
- Size effect refers to the outperformance of small cap securities over large cap ones. While research done in 1981 showed the existence of this effect, later research showed no supporting evidence.
- Value effect refers to stocks with low price to earnings (P/E) or low price to book (P/B), high dividend yields consistently outperforming stock with high P/E, P/B or low dividend yields. The presence of value effect challenges the semi-strong market efficiency as all the information to value stocks is publicly available and abnormal returns should not be possible. However, some researchers attribute the outperformance to higher risk in such stocks that is not appropriately captured in valuation models like CAPM.
- Other anomalies:
- Closed-end investment fund discounts: Closed-end investment funds offer ownership stake to subscribers at the inception of the fund and close the membership thereafter. Shares of such closed-end investment funds have typically traded at discount to their net asset value (NAV) per share (value of constituent securities net of fund liabilities divided by the number of shares outstanding). Various explanations offered include management fees, lack of investor control on timing of gain/loss realization etc. But the most plausible reason till date has been a lack of liquidity vs. publicly traded shares. Note it is not easy to make arbitrage profits on this anomaly, as transaction costs involved are too high (buying all the constituent securities and liquidating the fund) and the discount tends to narrow down over time.
- Earnings surprise refers to the piece of information in the earnings release of a company that was not anticipated by the investors and can impact company’s cash flows. Positive surprises lead to a rise in stock prices and vice versa. However, in a semi-strong market, the price adjustment should be quick enough. But studies show that the price adjustment sometimes continues well after the announcement. This lag can give the opportunity to investors earn abnormal returns using publicly available information. However, transaction costs and risks associated with such strategy need to be considered in determining the true return potential.
- IPOs (Initial Public Offerings) refer to the practice of offering a company stock to the public for the first time. Pricing of an IPO is a very important aspect as the investment bank selling the offering has to make sure that the issue is fully subscribed. This leads to prices being kept relatively lower. When the stock starts trading on the exchanges, the quantum of underpricing of the IPO is known (closing price on day one vs. the issue price). However, the performance of IPO in the subsequent periods has been found to be lagging the day one performance. This suggests that markets overprice the security initially. Some researchers, however, point out that the small size of IPO companies and statistics used may be resulting in such hypothesis.
- Predictability of returns on known economic factors: Some studies suggest stock returns may be related to knowing economic factors like inflation, interest rates, GDP growth rate etc. However, relationship alone can’t justify the tag of a market anomaly, as it will not be possible to earn abnormal returns just on knowing the fundamentals. Also, the relationship has to hold true over a long period of time.
Implications for investors:
Extracting abnormal returns from market anomalies has proved challenging for most investors because of the transaction costs involved. Further, most anomalies have turned out to be statistical biases than true correlations. So, while knowledge of potential anomalies is important, having a trading strategy based solely on anomalies may not be a very good idea.
Check your concepts:
(47.9) Which of the following statements is least accurate about the value effect anomaly?
(a) High dividend yielding stocks outperform low dividend yielding stocks
(b) High P/E stocks outperform low P/E stocks
(c) The presence of this anomaly challenges the semi-strong form market efficiency
(47.10) Which of the following market anomalies is most likely to challenge the weak form market efficiency?
(a) Value effect anomaly
(b) Momentum effect anomaly
(c) Size effect anomaly
Solutions:
(47.9) Correct Answer is B: According to the value effect anomaly, small P/E stocks outperform the high P/E stocks.
(47.10) Correct Answer is B: According to the momentum effect anomaly, one can earn excess return using technical analysis. Thus, it challenges the weak form market efficiency.
